GitLab

Moved all the sag related code to comfy_extras/nodes_sag.py

* First SAG test

* need to put extra options on the model instead of patcher

* no errors and results seem not-broken

* Use @ashen-uncensored formula, which works better!!!

* Fix a crash when using weird resolutions. Remove an unnecessary UNet call

* Improve comments, optimize memory in blur routine

* SAG works with sampler_cfg_function

The img2vid model is conditioned on clip vision output only which means
there's no CLIP model which is why I added a ImageOnlyCheckpointLoader to
load it. Note that the unClipCheckpointLoader can also load it because it
also has a CLIP_VISION output.

SDV_img2vid_Conditioning is the node used to pass the right conditioning
to the img2vid model.

VideoLinearCFGGuidance applies a linearly decreasing CFG scale to each
video frame from the cfg set in the sampler node to min_cfg.

SDV_img2vid_Conditioning can be found in conditioning->video_models
ImageOnlyCheckpointLoader can be found in loaders->video_models
VideoLinearCFGGuidance can be found in sampling->video_models

By adding a downscale to the unet in the first timesteps this node lets
you generate images at higher resolutions with less consistency issues.

This doesn't affect how percentages behave in the frontend but breaks
things if you relied on them in the backend.

percent_to_sigma goes from 0 to 1.0 instead of 1.0 to 0 for less confusion.

Make percent 0 return an extremely large sigma and percent 1.0 return a
zero one to fix imprecision.

This should make it much easier to write custom nodes with kdiffusion type
samplers.

This was broken because of recent changes so I fixed it and moved it from
the experiments repo.

Use the "lcm" sampler to sample them, you also have to use the
ModelSamplingDiscrete node to set them as lcm models to use them properly.

This allows changing the sampling parameters of the model (eps or vpred)
or set the model to use zsnr.